What to study for each 1Z0-1127-26 domain
Fundamentals of Large Language Models
This area covers how large language models work, including transformer architecture, tokens and embeddings, pretraining, and the decoding parameters that shape model output. It also covers prompt engineering techniques, fine-tuning approaches such as T-Few, and how LLM applications are evaluated.
Key skills
Transformer architecture, self-attention, tokens and embeddings, temperature and top-p, prompt engineering, few-shot prompting, fine-tuning, T-Few and PEFT, LLM evaluation
Using OCI Generative AI Service
This area covers Oracle's managed Generative AI service in depth: the pretrained foundation models from Cohere and Meta, playground and endpoint use, and the choice between on-demand inference and dedicated AI clusters. It also covers the inference, chat, and embedding APIs, custom model fine-tuning, and content moderation.
Key skills
Pretrained foundation models, Cohere and Llama models, dedicated AI clusters, on-demand endpoints, inference and chat APIs, embedding models, custom model fine-tuning, content moderation, model lifecycle
Implement RAG using OCI Generative AI Service
This area covers building retrieval-augmented generation solutions on OCI, from embedding documents and chunking strategies to vector search in Oracle Database 23ai. It also covers knowledge bases, the OCI Generative AI Agents service, and grounding answers with citations.
Key skills
Embeddings, chunking strategies, AI Vector Search, Oracle Database 23ai, knowledge bases, Generative AI Agents, retrieval ranking, grounding and citations, hybrid search
Building and Deploying LLM Applications
This area covers assembling production LLM applications on OCI, including LangChain integration, OCI Data Science for custom models, and IAM policies that secure GenAI resources. It also covers observability, scaling, rate limits, and cost controls for deployed applications.
Key skills
LangChain, OCI Data Science, IAM policies, API security, observability and tracing, scaling and rate limits, agent orchestration, cost controls, testing GenAI applications